All projects described in this proposal will require access to dedicated biostatistical consulting and computing support. Individual project leaders will need access to standard biostatistical consulting support, and overall project support to analyze expression and sequence data, and correlate these with clinical data for discovery. This Core will have three specific aims that will be accomplished by the application of advanced statistical data analysis algorithms. 1. We will explore the relevance of individual confirmed mutations for the clinical outcomes of our discovery group of 47 AML patients. Statistical methods identical to the ones used by the CALGB statistical group, and in collaboration with that group will be performed. These methods include Fisher's Exact Test, Wilcoxon rank-sum test, Kaplan-Meier, and Cox proportional hazard's models. 2. We will determine whether individual mutations predict signature expression profiles. The statistical methods used to perform these studies include hierarchical and k-means cluster analysis, principal components analysis, and Mantel correlation. This kind of study has already been successfully performed for t(15;17) positive patients with M3 AML, but our analysis of profiles from tumors containing the FLT3 mutation did not predict expression profiles. The identification of profiles predicted by mutations will provide a rapid diagnostic tool that would validate mutational analysis, and also perhaps provide clues about pathways affected by a specific mutation. 3. We will determine whether specific RNA profiles are predictive of clinical outcomes. The studies used in this Aim will be similar to that of Aim 2. An example of this has already been provided in the analysis of 8 M2 and M3 patients who failed to respond to initial therapy, in which a group of differentially expressed genes was identified. The methodologies used to analyze the data sets are already in place in the Department of Medicine's Biostatistical Consulting Center, which is directed by Dr. Shannon and where the work of the Core will be done, and they will be coordinated with the CALGB Statistical Center for consistency of approach.